scholarly journals Random-effect based test for multinomial logistic regression: choice of the reference level and its impact on the testing

2021 ◽  
Author(s):  
Qianchuan He ◽  
Yang Liu ◽  
Meiling Liu ◽  
Michael C. Wu ◽  
Li Hsu
Keyword(s):  
Logistic Regression ◽  
Genetic Variants ◽  
Multinomial Logistic Regression ◽  
Random Effect ◽  
Score Test ◽  
Binary Logistic Regression ◽  
Reference Level ◽  
Invariance Property ◽  
Recent Effort ◽  
Numerical Studies

ABSTRACTRandom-effect score test has become an important tool for studying the association between a set of genetic variants and a disease outcome. While a number of random-effect score test approaches have been proposed in the literature, similar approaches for multinomial logistic regression have received less attention. In a recent effort to develop random-effect score test for multinomial logistic regression, we made the observation that such a test is not invariant to the choice of the reference level. This is intriguing because binary logistic regression is well-known to possess the invariance property with respect to the reference level. Here, we investigate why the multinomial logistic regression is not invariant to the reference level, and derive analytic forms to study how the choice of the reference level influences the power. Then we consider several potential procedures that are invariant to the reference level, and compare their performance through numerical studies. Our work provides valuable insights into the properties of multinomial logistic regression with respect to random-effect score test, and adds a useful tool for studying the genetic heterogeneity of complex diseases.

scholarly journals Genetic association tests in family samples for multi-category phenotypes

BMC Genomics ◽  
2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Shuai Wang ◽  
James B. Meigs ◽  
Josée Dupuis
Keyword(s):  
Logistic Regression ◽  
Statistical Methods ◽  
Error Rate ◽  
Type I Error ◽  
Multinomial Logistic Regression ◽  
Score Test ◽  
Type I ◽  
Type I Error Rate ◽  
Ordinal Traits ◽  
Study Designs

Abstract Background Advancements in statistical methods and sequencing technology have led to numerous novel discoveries in human genetics in the past two decades. Among phenotypes of interest, most attention has been given to studying genetic associations with continuous or binary traits. Efficient statistical methods have been proposed and are available for both types of traits under different study designs. However, for multinomial categorical traits in related samples, there is a lack of efficient statistical methods and software. Results We propose an efficient score test to analyze a multinomial trait in family samples, in the context of genome-wide association/sequencing studies. An alternative Wald statistic is also proposed. We also extend the methodology to be applicable to ordinal traits. We performed extensive simulation studies to evaluate the type-I error of the score test, Wald test compared to the multinomial logistic regression for unrelated samples, under different allele frequency and study designs. We also evaluate the power of these methods. Results show that both the score and Wald tests have a well-controlled type-I error rate, but the multinomial logistic regression has an inflated type-I error rate when applied to family samples. We illustrated the application of the score test with an application to the Framingham Heart Study to uncover genetic variants associated with diabesity, a multi-category phenotype. Conclusion Both proposed tests have correct type-I error rate and similar power. However, because the Wald statistics rely on computer-intensive estimation, it is less efficient than the score test in terms of applications to large-scale genetic association studies. We provide computer implementation for both multinomial and ordinal traits.

scholarly journals Genetic association tests in family samples for multi-category phenotypes

2021 ◽  
Author(s):  
Shuai Wang ◽  
James Meigs ◽  
Josee Dupuis
Keyword(s):  
Logistic Regression ◽  
Statistical Methods ◽  
Error Rate ◽  
Type I Error ◽  
Multinomial Logistic Regression ◽  
Score Test ◽  
Type I ◽  
Type I Error Rate ◽  
Ordinal Traits ◽  
Study Designs

Abstract Background Advancements in statistical methods and sequencing technology have led to numerous novel discoveries in human genetics in the past two decades. Among phenotypes of interest, most attention has been given to studying genetic associations with continuous or binary traits. Efficient statistical methods have been proposed and are available for both type of traits under different study designs. However, for multinomial categorical traits in related samples, there is a lack of widely used efficient statistical methods and software. Results We propose an efficient score test to analyze a multinomial trait in family samples, in the context of genome-wide association/sequencing studies. An alternative Wald statistic is also proposed. We also extend the methodology to be applicable to ordinal traits. We performed extensive simulation studies to evaluate the type-I error of the score test, Wald test compared to the multinomial logistic regression for unrelated samples, under different allele frequency and study designs. We also evaluate the power of these methods. Results show that both the score and Wald tests have well-controlled type-I error rate, but the multinomial logistic regression has inflated type-I error rate when applied to family samples. We illustrated the application of the score test with an application to the Framingham Heart Study to uncover genetic variants associated with diabesity, a multi-category phenotype. Conclusion Both proposed tests have correct type-I error rate and similar power rate. However, because the Wald statistics rely on computer intensive estimation, it is less efficient than the score test in terms of applications to large-scale genetic association studies. We provide computer implementation for both multinomial and ordinal traits.

scholarly journals Domestic and International Remittances and Food Security in Sub-Saharan Africa

2020 ◽  
Vol 5 (1) ◽  
pp. 37-54
Author(s):  
Narges Ebadi ◽  
Davod Ahmadi ◽  
Hugo Melgar-Quiñonez
Keyword(s):  
Logistic Regression ◽  
Food Security ◽  
Food Insecurity ◽  
Multinomial Logistic Regression ◽  
Binary Logistic Regression ◽  
Sub Saharan Africa ◽  
International Remittances ◽  
Food Security Status ◽  
Sub Saharan ◽  
The Impact

The amount of remittances to developing counties, defined as the flow of monetary and non-monetary goods, has increased globally and has surpassed the amount of money spent on foreign aid in these developing countries. The impact of remittances on households’ purchasing power has been studied; however, its link to food security status is yet to be explored. This paper quantitatively analyses the relationship between food security status (measured using the Food Insecurity Experience Scale) and the receipt of domestic/ international or both remittances on households in sub- Saharan Africa. Data are derived from the Gallup World Poll from the years 2014-2017. Multinomial logistic regression models and binary logistic regression analyses were conducted to analyze the data. Results showed that remittance recipients had significantly higher household incomes (especially if the remittance was coming internationally and domestically), lived with significantly more household members (7 or more members), and were more likely to be separated (including divorced or widowed). Households that received domestic remittances had significantly higher odds of being food insecure than households receiving no remittances. Conversely, households receiving remittances internationally or a combination of domestic and international remittances had significantly lower odds of food insecurity compared to non-receivers. This study found that receiving remittances affect the food security status of people living in SSA countries. 

Social networks and their effects on the choice of contraceptive use and method in Bangladesh

2019 ◽  
Vol 13 (1) ◽  
pp. 58-71 ◽  
Author(s):  
Md Shahidul Islam
Keyword(s):  
Social Networks ◽  
Logistic Regression ◽  
Social Network ◽  
Contraceptive Use ◽  
Multinomial Logistic Regression ◽  
Binary Logistic Regression ◽  
Modern Method ◽  
Modern Contraceptive ◽  
Content Type ◽  
Modern Contraceptive Method

Purpose The purpose of this paper is to examine the effects of social networks on contraceptive adoption in Bangladesh. Design/methodology/approach Data were collected from 430 couples; the age of men was between 15 and 49 years. χ2 test was applied to test the association between independent variables and current contraceptive use. Binary logistic regression was applied to examine the effects of social network on contraceptive use, and multinomial logistic regression was applied to examine the effect of social network on the choice of method. Findings Results from binary logistic regression reveal that social networks of both men (OR=2.71, 95% CI=1.371−5.354) and women (OR= 3.597, 95% CI=1.754−7.380) had a strong positive effect on current contraceptive use. The result from multinomial logistic regression also shows that men’s social network (OR= 2.74, 95% CI= 1.356−5.548) and women’s social network (OR=4.165,95% CI=1.958−8.860) were also associated with choosing a modern contraceptive method. Originality/value Social networks have a significant effect on current contraceptive use and modern method choice in Bangladesh. A social network approach should be included in family planning program in Bangladesh.

scholarly journals Correlates of patterns of cannabis use, abuse and dependence: evidence from two national surveys in Ireland

2021 ◽  
Author(s):  
Seán R Millar ◽  
Deirdre Mongan ◽  
Claire O’Dwyer ◽  
Jean Long ◽  
Bobby P Smyth ◽  
...  
Keyword(s):  
Logistic Regression ◽  
Multinomial Logistic Regression ◽  
Binary Logistic Regression ◽  
Cannabis Use ◽  
Educational Levels ◽  
Factors Associated ◽  
Diagnostic And Statistical Manual ◽  
Private Households ◽  
National Drug ◽  
Younger Age

Abstract Background Knowledge of factors relating to patterns of cannabis use is important for informing drug policy. This study determined factors associated with recent and current cannabis use. In addition, we explored factors related to having a cannabis use disorder (CUD)—defined using the Diagnostic and Statistical Manual of Psychiatric Disorders—among current users. Methods We analyzed data from Ireland’s 2010–11 and 2014–15 National Drug Prevalence Surveys, which recruited 5134 and 7005 individuals respectively, aged 15 years or more, living in private households. Multinomial logistic regression was used to identify factors associated with recent (last year) and current (last month) cannabis use compared to experiential use. Binary logistic regression was used to determine factors related to CUD among current users. Results The weighted prevalence of experiential cannabis use was 18.3%, with 3.0% and 3.3% of participants indicating recent or current use, respectively; 41.3% of current users indicated having a CUD. Factors associated with both recent or current cannabis use included younger age, not being married or cohabiting, having no dependent children and current use of tobacco or alcohol. Male gender, younger age and lower educational levels were significantly related to CUD among current users. Conclusions Males, adolescents/young adults and individuals with lower educational levels are more likely to be current users of cannabis and are at a greater risk of having a CUD. Health professionals should be aware of these factors to improve detection and prevention of CUD.

scholarly journals Modeling of child labour exploitation status in Indonesia using multilevel binary logistic regression

2021 ◽  
Vol 36 ◽  
pp. 01008
Author(s):  
Liza Kurnia Sari ◽  
Lissa Octavia Wardana
Keyword(s):  
Logistic Regression ◽  
Random Effect ◽  
Household Head ◽  
Binary Logistic Regression ◽  
Working Hours ◽  
Child Labour ◽  
Individual Level ◽  
Exploitation Status ◽  
The Individual ◽  
The Right

The Indonesian constitution recognizes guarantees the right of the child to rest and leisure, to engage in play, and recreational activities appropriate to the age of the child so that they should not be working. Employers are also prohibited to employ children. However, many children come to work because of poverty, even though child labour is close to exploitation. Theoretically, individual and contextual factors affect the exploitation status of child labour. This study aims to analyze the variables that influence the exploitation of child labour in Indonesia based on data from the National Socio-Economic Survey (Susenas) in 2018. The random effect test shows that there are differences between regency/municipality so that multilevel binary logistic regression performs better than one level binary logistic regression. More than 80 percent of child labourers are exploited in terms of education and working hours. Variables that significantly influence the exploitation status of child labour at the individual level are gender, the occupation sector of child labour, and the occupation sector of the household head. Meanwhile, poverty rates and mean years of schooling significantly influence the exploitation status of child labour at the regional level.

scholarly journals Exploring the Influencing Factors for Infant Mortality: A Mixed-Method Study of 24 Developing Countries Based on Demographic and Health Survey data

2021 ◽  
Author(s):  
Md. Akhtarul Islam ◽  
Tarana Tabassum ◽  
Mohammad Ali Moni
Keyword(s):  
Developing Countries ◽  
Logistic Regression ◽  
Infant Mortality ◽  
Regression Model ◽  
Mixed Method ◽  
Logistic Regression Model ◽  
Random Effect ◽  
Binary Logistic Regression ◽  
Forest Plot ◽  
Binary Logistic Regression Model

Objective: This study aimed to discover the prevalence of infant mortality and to assess how different factors influence infant mortality in 24 developing countries by utilizing the latest DHS data. Methods: This study used a mixed-method design to assemble cross-sectional studies to integrate data from 24 other countries due to a widening perspective of infant mortality. Most recent available DHS data of 24 different developing countries from the year 2013 to 2019 was used to conduct the study. Descriptive analysis, binary logistic regression model, random-effect meta-analysis, and forest plot have been used for the final analyses. Results: Binary logistic regression model revealed for Bangladesh that, higher education level of fathers (OR: 0.344, 95% CI: 0.147; 0.807), being 2nd born or above order infant (OR: 0.362, 95% CI: 0.248, 0.527), taking ANC (OR: 0.271, 95% CI: 0.192; 0.382 for 1-4 visits), taking PNC (OR: 0.303, 95% CI: 0.216; 0.425) were statistically significant determinants of lowering infant death. While carrying multiple fetus (OR: 6.634, 95% CI: 3.247; 13.555) was exposed as a risk factor of infant mortality. Most significant factors influencing infant mortality for all 24 developing countries were number of fetus (OR: 0.193, 95% CI: 0.176; 0.213), taking ANC (OR: 0.356, 95% CI: 0.311; 0.407) and taking PNC (OR: 0.302, 95% CI: 0.243; 0.375). Conclusion In this study, the number of the fetus, taking ANC and PNC, was the most significant factor affecting the risk of infant mortality in developing countries. So, anticipation and control projects ought to be taken in the field in regard to these hazard factors.

scholarly journals Association of Dietary n3 and n6 Fatty Acids Intake with Hypertension: NHANES 2007–2014

Nutrients ◽  
2019 ◽  
Vol 11 (6) ◽  
pp. 1232 ◽  
Author(s):  
Jiahao Chen ◽  
Baoqi Sun ◽  
Dongfeng Zhang
Keyword(s):  
Fatty Acids ◽  
Logistic Regression ◽  
Multinomial Logistic Regression ◽  
Binary Logistic Regression ◽  
Cross Sectional Study ◽  
Newly Diagnosed ◽  
Cross Sectional ◽  
Dietary Recall ◽  
Health And Nutrition ◽  
Adjusted Model

We conducted this cross-sectional study in the American general population to explore the association of dietary n3 and n6 fatty acids intake and the risk of hypertension. We used data from the National Health and Nutrition Examination Survey (NHANES) 2007–2014 in this study. We obtained dietary n3 and n6 fatty acids data through two 24 h dietary recall interviews and n3, n6 fatty acids intake were adjusted by weight. We defined hypertension as now taking prescribed medicine for hypertension or blood pressure above 130/80 mmHg. We applied binary logistic regression, multinomial logistic regression, and restricted cubic spline to evaluate the associations of dietary n3 and n6 fatty acids intake with hypertension. A total of 18,434 participants were included in this study. In the multivariate-adjusted model 2, the odds ratios (ORs) with 95% confidence interval (CI) of hypertension were 0.58 (0.49–0.68), 0.53 (0.45–0.63), and 0.92 (0.80–1.06) for the highest versus the lowest tertile of dietary n3, n6 fatty acids intake and n6:n3 ratio, respectively. Further excluded participants with hypertension history, the ORs with 95% CI of newly diagnosed hypertension were 0.60 (0.50–0.73), 0.52 (0.43–0.62), and 0.95 (0.79–1.14) for the highest versus lowest tertile of dietary n3, n6 fatty acids intake and n6:n3 ratio, respectively. Dose-response analyses showed that the risk of hypertension was associated with dietary n3 and n6 fatty acids intake. Our study suggested that dietary n3 and n6 fatty acids intake were inversely associated with the risk of hypertension in US adults.

scholarly journals Factors Associated With the Intention to Begin Physical Activity Among Inactive Middle-Aged and Older Adults

2021 ◽  
pp. 109019812110308
Author(s):  
Ariane S. Massie ◽  
Heather Johnston ◽  
Daniel Sibley ◽  
Brad A. Meisner
Keyword(s):  
Physical Activity ◽  
Older Adults ◽  
Logistic Regression ◽  
Transtheoretical Model ◽  
Multinomial Logistic Regression ◽  
Binary Logistic Regression ◽  
Later Life ◽  
Canadian Community Health Survey ◽  
Middle Aged ◽  
Psychological Predictor

Factors that affect physical activity (PA) behavior change are well established. Behavioral intention is a strong psychological predictor of behavior; however, there is less research on the factors that affect the intention to increase PA participation specifically, especially among adults in mid and later life who are inactive. Using data from the Canadian Community Health Survey, which was informed by the transtheoretical model (TTM), this study investigated the relationships between a range of demographic and biopsychosocial factors with the intention to become physically active among 1,159 inactive adults aged 40 years and older. Comparisons were made between participants reporting the intention to begin PA in the next 30 days (TTM Preparation; n = 610), 6 months (TTM Contemplation; n = 216), or not at all (TTM Precontemplation; n = 333). First, multinomial logistic regression identified age, sex, ethnicity, education, restriction of activities, self-perceived health, and community belonging as factors significantly associated with 30-day PA intention, while age and ethnicity were significantly associated with 6-month PA intention, compared with those reporting no intention. Second, binary logistic regression revealed that education was the only factor that differentially associated with intention timeframe as participants with lower levels of education were less likely to report PA intention in 30 days compared with 6 months. Findings demonstrate key demographic, biopsychosocial, and temporal factors that warrant consideration for tailored PA promotion programs that aim to effectively address the constraints and barriers that negatively influence PA intention among middle-aged and older adults.

Sign in / Sign up

Export Citation Format

Share Document